summaryrefslogtreecommitdiffstats
path: root/cmake/modules/Findxfs.cmake (unfollow)
Commit message (Collapse)AuthorFilesLines
2019-02-22civetweb: pull up to ceph-masterAbhishek Lekshmanan1-0/+0
Contains the following fixes - absolute url fixes - ssl socket closing fixes Signed-off-by: Abhishek Lekshmanan <abhishek@suse.com>
2019-02-22doc: Remove value 'mon_osd_max_split_count' from the docsKai Wagner1-11/+0
Signed-off-by: Kai Wagner <kwagner@suse.com>
2019-02-22cmake: fix syntax error of set()Kefu Chai1-1/+1
Signed-off-by: Kefu Chai <kchai@redhat.com>
2019-02-22install-deps,do_freebsd: disable amqp supportKefu Chai2-1/+1
per Willem's request. Signed-off-by: Kefu Chai <kchai@redhat.com>
2019-02-22rpm: disable amqp endpoint on SLE/OpenSUSEKefu Chai1-0/+7
as librabbitmq-devel is not packaged on them yet. Signed-off-by: Kefu Chai <kchai@redhat.com>
2019-02-22cmake: remove amqp_mock from UNITTEST_LIBSKefu Chai2-18/+22
this mock library is only used by rgw tests. Signed-off-by: Kefu Chai <kchai@redhat.com>
2019-02-21rpm: build __pycache__ directories on RHEL 8Ken Dreyer1-0/+3
In RHEL 8, rpmbuild does not automatically generate __pycache__ directories. Generate them explicitly with the %py_byte_compile macro. Signed-off-by: Ken Dreyer <kdreyer@redhat.com>
2019-02-21rpm: fix pyOpenSSL requirement for RHEL 8Ken Dreyer1-1/+4
On RHEL 7, this package is named "pyOpenSSL", but on RHEL 8, Fedora, and SUSE, it is named "python3-pyOpenSSL". Prior to this change, the RHEL conditional was applying to both RHEL 7 and 8, so we would fail to resolve the "pyOpenSSL" package dependency on RHEL 8. Update the RPM conditionals to use "pyOpenSSL" on RHEL 7 and "python3-pyOpenSSL" on all other distributions. Signed-off-by: Ken Dreyer <kdreyer@redhat.com>
2019-02-21rpm: fix python dependencies on RHEL 8Ken Dreyer1-1/+1
On RHEL 8, the python34-* and Cython packages are named identically to Fedora (eg "python3-Cython"). Move the Python 2 "Cython" and python34-* BuildRequires lines inside a RHEL 7 conditional, so RHEL 8 falls through to the Fedora package names. Signed-off-by: Ken Dreyer <kdreyer@redhat.com>
2019-02-21unittest_auth: add unit tests for AuthRegistry con_mode selectionSage Weil3-0/+255
Signed-off-by: Sage Weil <sage@redhat.com>
2019-02-21auth/AuthRegistry: register as an observerSage Weil2-8/+21
Signed-off-by: Sage Weil <sage@redhat.com>
2019-02-21auth/AuthRegistry: clear output vectorsSage Weil1-0/+9
Signed-off-by: Sage Weil <sage@redhat.com>
2019-02-21msg/async: guard protocol Interceptor with preprocessor variableRicardo Dias4-0/+20
Signed-off-by: Ricardo Dias <rdias@suse.com>
2019-02-21cmake: add UNIT_TESTS_BUILD preprocessor var based on WITH_TESTS cmake varRicardo Dias2-0/+4
Signed-off-by: Ricardo Dias <rdias@suse.com>
2019-02-21pybind: auto encode/decode cstrKefu Chai2-2/+2
assuming all string in the cythonized string use the legacy python2 string type, instead of using the unicode, and assuming all c strings are ASCII, so we don't need encode/decode them when coerce from/to C strings. this helps when generating py3 bindings Signed-off-by: Kefu Chai <kchai@redhat.com>
2019-02-21pybind: set language_level for cythonize explicitlyKefu Chai4-1/+6
Compiling rbd.pyx because it changed. [1/1] Cythonizing rbd.pyx /usr/lib/python2.7/dist-packages/Cython/Compiler/Main.py:367: FutureWarning: Cython directive 'language_level' not set, using 2 for now (Py2). This will change in a later re lease! File: /var/ssd/ceph/src/pybind/rbd/rbd.pyx tree = Parsing.p_module(s, pxd, full_module_name) this warning is raised by cython 0.29.2 Signed-off-by: Kefu Chai <kchai@redhat.com>
2019-02-21vstart: make beast as the default frontend for rgwAbhishek Lekshmanan2-2/+3
mstart also features an overrideable $RGW_FRONTEND which defaults to beast Signed-off-by: Abhishek Lekshmanan <abhishek@suse.com>
2019-02-21Revert "qa: update test_envlibrados_for_rocksdb.sh for libradospp split"Jason Dillaman2-71/+0
This reverts commit 0475dfdea315e5208abc201355523c8dab78469f. Signed-off-by: Jason Dillaman <dillaman@redhat.com>
2019-02-21doc/librados: explicitly mention that the C++ API is not stableJason Dillaman1-0/+4
Signed-off-by: Jason Dillaman <dillaman@redhat.com>
2019-02-21ceph.spec: force use of upgrade devtoolset-gcc under RHEL 7Jason Dillaman1-1/+1
There is a known crash under C++17 mode that requires an updated devtoolset to address. See https://bugzilla.redhat.com/show_bug.cgi?id=1616223 Signed-off-by: Jason Dillaman <dillaman@redhat.com>
2019-02-21librados: add symbol versioning to the C++ APIJason Dillaman51-213/+153
The future goal would be to change the version for each Ceph major release to ensure C++ applications will need to be recompiled against the librados C++ API since we don't guarentee ABI stability. Fixes: https://tracker.ceph.com/issues/38177 Signed-off-by: Jason Dillaman <dillaman@redhat.com>
2019-02-21doc/rbd: fix options of iscsiadm commandykcowrebbaj1-1/+1
Signed-off-by: ZhuJieWen <154667715@qq.com>
2019-02-21rgw,cmake: add option "WITH_RADOSGW_AMQP_ENDPOINT"Kefu Chai5-20/+56
WITH_RADOSGW_AMQP_ENDPOINT is enabled by default. Signed-off-by: Kefu Chai <kchai@redhat.com>
2019-02-21mgr/ansible: Add/remove hostsJuan Miguel Olmo Martínez3-78/+312
- Implementation of operations add/rm hosts - Improvement of error management in Ansible Runner Service Client - Addressed @sebastian-philipp suggestions and findings - Remove the <role> parameter of hosts operations. Now all the hosts are placed by default in a specific Ansible Inventory group called <orchestrator>. Discussion on-going about to keep Host management API endpoint or not. If hosts management is removed, further modifications will be needed to: - Keep updated the <host inventory> in the orchestrator - Manage hosts groups when operations are executed - Adapted new commands (host add/rm) to use to the new <orchestrator_cli> syntax - Improved error management - Added return information to Add/remove host methods docstring - Removed redundant exceptions following @sebastian-philipp suggestion - Removed duplicated code - Removed extra space - Fixed unit tests - Improved error management in launch playbook - removed unused param in docstring (orchestrator.py) Signed-off-by: Juan Miguel Olmo Martínez <jolmomar@redhat.com>
2019-02-21cmake: find_package(RabbitMQ) instead of just checking its headerKefu Chai3-11/+26
rabbitmq-c is not a header-only library. so we need to ensure that the library is also around before linking against it. in this change: * move the detection down to where librabbitmq is used * add FindRabbitMQ.cmake for find_package(RabbitMQ) * s/rabbitmq/RabbitMQ::RabbitMQ/ as the latter is an alias library target, which has more information attached to it. Signed-off-by: Kefu Chai <kchai@redhat.com>
2019-02-21install-deps.sh: do not install librabbitmq-dev manuallyKefu Chai1-1/+0
it will be installed by `mk-build-deps --install`. Signed-off-by: Kefu Chai <kchai@redhat.com>
2019-02-21librbd: execute image pre-removal validation before moving image to trashJason Dillaman2-5/+31
Fixes: http://tracker.ceph.com/issues/38404 Signed-off-by: Jason Dillaman <dillaman@redhat.com>
2019-02-21librbd: split image pre-removal logic to separate state machineJason Dillaman8-584/+870
Signed-off-by: Jason Dillaman <dillaman@redhat.com>
2019-02-21msg/async/rdma: parse IBSYNMsg.lid as hex when receiving messageLiuPeng1-1/+1
function send_msg encode struct IBSYNMsg to a string, and recv_msg parse the string. In the function send_msg, im.lid (unsigned short int) is formatted into a string with format string %04x, so, recv_msg should read im.lid from the string with format string %hx instead of %hu (unsigned short). Fixes: https://tracker.ceph.com/issues/38391 Signed-off-by: Peng Liu <liupeng37@baidu.com>
2019-02-21mgr: add per pool scrub commandsxie xingguo3-0/+88
Signed-off-by: xie xingguo <xie.xingguo@zte.com.cn>
2019-02-21osdc/Objecter: invalidate crcs on preallocated rx buffersIlya Dryomov2-0/+31
Both simple and async messengers use c_str() when copying the data from the socket into the receive buffer, going behind bufferlist's back. If the receive buffer is preallocated, we need to invalidate its crc cache by hand to avoid possible data crc mismatches on the client side. Fixes: https://tracker.ceph.com/issues/38416 Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2019-02-21mon/AuthMonitor: provide auth_lock-safe _assign_global_id()Sage Weil2-17/+44
Break assign_global_id() into two parts: - _assign_global_id(), called under auth_lock, that does nothing but assign an id (or fail). This makes use of the num_mon and num_rank values, which means it only works while we are in quorum. - increase_max_global_id(), if it looks like we should. Also, adjust _should_increase_max_global_id() to also work under auth_lock since we are looking at last_allocated_id. All last_allocated_id users are not protected by auth_lock. Fixes: http://tracker.ceph.com/issues/38372 Signed-off-by: Sage Weil <sage@redhat.com>
2019-02-21osd/PG: fix get debug refcount lineSage Weil1-1/+1
Signed-off-by: Sage Weil <sage@redhat.com>
2019-02-21mon/AuthMonitor: provide AuthMonitor with mon count and rank *only* while activeSage Weil3-29/+58
Provide AuthMonitor with the mon count and rank under the mon->auth_lock. Only set these values while the monitor is active (and the mon is allowed to authenticate). When we start an election or bootstrap, clear these values. Note that we aren't using them yet! Signed-off-by: Sage Weil <sage@redhat.com>
2019-02-20librbd: drop extra check for image watchers from trash move APIJason Dillaman6-37/+11
Signed-off-by: Jason Dillaman <dillaman@redhat.com>
2019-02-20librbd: attempt to restore deleted image from trash on known failuresJason Dillaman1-0/+6
If the image has watchers, snapshots, or group links, restore the image from the trash back to the image directory since data removal hasn't yet started. Signed-off-by: Jason Dillaman <dillaman@redhat.com>
2019-02-20rbd: don't hide partially deleted images in long image listJason Dillaman1-5/+4
Signed-off-by: Jason Dillaman <dillaman@redhat.com>
2019-02-20librbd: include in-progress image removals in image listJason Dillaman13-24/+159
Exclude them from the trash list to prevent double-counting them. This mimics the previous behavior for handling interrupted image removals. Signed-off-by: Jason Dillaman <dillaman@redhat.com>
2019-02-20librbd: fixed some edge cases when moving images to trash upon removeJason Dillaman10-146/+137
Signed-off-by: Jason Dillaman <dillaman@redhat.com>
2019-02-20librbd: skip mirror image verification when pruning rbd-mirror watchersJason Dillaman3-46/+2
When an image is moved to the trash, mirroring is disabled which will remove the image id from the mirroring directory. If you then quickly attempt to remove the image from the trash, it might fail if the rbd-mirror daemon hasn't closed the image since it fails to filter out rbd-mirror daemon watchers against images with mirroring disabled. Signed-off-by: Jason Dillaman <dillaman@redhat.com>
2019-02-20rbd: move image to trash as a first stepMahati Chamarthy9-55/+165
... while removing an Image Fixes: http://tracker.ceph.com/issues/24226 Signed-off-by: Mahati Chamarthy <mahati.chamarthy@intel.com>
2019-02-20librados: add symbol versioning to the C APIJason Dillaman8-615/+984
Signed-off-by: Jason Dillaman <dillaman@redhat.com>
2019-02-20librados: revert librados3/libradoscc back to librados2Jason Dillaman42-181/+101
For backwards compatibility and upgrade reasons, the librados2 API needs to be preserved and it needs to continue to be compatible with dependent libraries like librbd1. Signed-off-by: Jason Dillaman <dillaman@redhat.com>
2019-02-20librbd: CopyupRequest::m_pending_copyups doesn't need to be atomicIlya Dryomov2-2/+2
It's protected by CopyupRequest::m_lock mutex. Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2019-02-20librbd: don't clone IoCtx in send_copyup()Ilya Dryomov2-5/+2
Calling aio_operate() with the explicit snap_seq and snaps doesn't mutate IoCtx, so no reason to clone it. Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2019-02-20librbd: send copyup and hint only onceIlya Dryomov1-12/+22
In the deep-copy case, don't send copyup and hint twice (once with the blank and once with the current snapshot context). Preserve the workaround for compare-and-write added in commit f6db9b8027b6 ("librbd: copyup state machine needs to handle empty write ops"). Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2019-02-20common/options: add ms_mon_client_modeSage Weil3-3/+24
The server now has a list of *allowed* modes (no ordering) and the clients have a list of modes in order of preference. Since we want everything connecting to the mon to be secure by default (think: ceph auth set ...), we need a separate option to ensure we prefer secure mon connections from the CLI etc. Signed-off-by: Sage Weil <sage@redhat.com>
2019-02-20auth: push con_mode selection into AuthRegistry helperSage Weil4-15/+33
Signed-off-by: Sage Weil <sage@redhat.com>
2019-02-20mgr/orch/ssh: add qa testNoah Watkins2-0/+44
Signed-off-by: Noah Watkins <noahwatkins@gmail.com>
2019-02-20qa: install/exclude ceph-mgr-ssh packageNoah Watkins12-0/+12
Signed-off-by: Noah Watkins <noahwatkins@gmail.com>
generated by cgit v1.2.3 (git 2.39.1) at 2025-01-19 04:21:58 +0100